Modeling Particulate Emissions in Europe. A Framework to Estimate Reduction Potential and Control Costs

(1)

International Institute for Applied Systems Analysis Schlossplatz 1 • A-2361 Laxenburg • Austria Telephone: (+43 2236) 807 342 • Fax: (+43 2236) 71313 E-mail: publications^@iiasa.ac.at • Internet: www.iiasa.ac.at

Interim Report IR-02-076

Modelling Particulate Emissions in Europe

A Framework to Estimate Reduction Potential and Control Costs

Zbigniew Klimont, Janusz Cofala, Imrich Bertok, Markus Amann, Chris Heyes and Frantisek Gyarfas

Approved by

Markus Amann (amann@iiasa.ac.at) Leader, Transboundary Air Pollution

(2)

(3)

Acknowledgments

The financial support received from the Umweltbundesamt Berlin is gratefully acknowledged.

The authors want to thank Rainer Remus, Matthias Tappe, Gunnar Gholisch and Bernd Schärer from the Umweltbundesamt Berlin for their assistance in conducting the study and for providing recent information on PM emissions from a series of ongoing German studies. Furthermore, the authors express their sincere thanks to Les White from White Associates (UK), Jozef Pacyna from the Norwegian Institute for Air Research (NILU), AEA Technology, Helen ApSimon and Teresa Gonzalez from Imperial College, London (UK), Helen Dunn from the UK Department of Environment, Food and Rural Affairs (DEFRA), Lena Lillieblad (ALSTOM Power Environmental Systems AB, Växjö, Sweden), Sten Maartmann (Sweden), Centre Interprofessionnel Technique d’Etudes de la Pollution Atmospherique (CITEPA), University of Stuttgart - Institute of Energy Economics and the Rational Use of Energy (IER) and Jan Berdowski, Antoon Visschedijk and Tinus Pulles from The Netherlands Organisation for Applied Scientific Research (TNO).

(8)

(9)

Abstract

This paper presents the extension of the Regional Air Pollution Information and Simulation (RAINS) model that addresses present and future emissions of fine particulates in Europe, the potential for controlling these emissions and the costs of such emission reductions. Together with the existing modules dealing with the emissions of the precursor emissions of secondary aerosols such as sulphur dioxide (SO2), nitrogen oxides (NOx), ammonia (NH3) and volatile organic compounds (VOC), this extension enables the comparison of the potentials and costs for controlling primary emissions of fine particles with those of secondary aerosols and to find cost- minimal approaches for reducing ambient levels of particulate matter.

The emissions of particulate matter (PM) in the RAINS model are calculated for three different size classes: the fine fraction (PM2.5), the coarse fraction (PM10 - PM2..5) and large particles (PM_>10 µm). Summed up, these three fractions represent total suspended particles (TSP).

Fine particles are emitted from a large number of sources with large differences in their technical and economic properties. The methodology distinguishes 392 source categories for stationary energy combustion, industrial processes, mobile sources and agriculture. For each of these sectors, the study explores the applicable options for reducing PM emissions, their efficiency and their costs.

Emissions characteristics of the individual sectors are strongly determined by country-specific conditions. The methodology estimates emission control costs of standard technologies under the specific conditions characteristic for the various European countries. Based on the assumption of the general availability of control technologies with equal technical properties and costs, a number of country-specific circumstances (level of technological advancement, installation size distribution, labor costs, etc.) are used to estimate the costs for the actual operation of pollution control equipment.

For the individual source sectors, emissions are estimated based on statistical information on economic activity and emission factors that reflect hypothetical emissions if no control measures were applied. These emission factors were taken from the literature and were, to the maximum possible extent, adapted to the country-specific conditions. Actual emissions are calculated taking into account the application of emission control measures in a given sector, for which also costs are estimated.

The methodology was implemented for all European countries, covering the period from 1990 to 2010. At an aggregated level, estimates for past years (1990, 1995) correspond well with other national and international inventories. However, discrepancies are found for some detailed results for individual sectors and activities, and more work will be necessary to clarify them.

This preliminary implementation suggests for Europe a 50 percent decline of primary emissions of fine particles between 1990 and 1995, mainly due to the economic restructuring in central

(10)

are expected to be 60 percent below the level of 1990. However, less improvement is expected for the health-relevant fraction of fine particles (PM2.5).

It needs to be emphasized that these preliminary estimates are still associated with considerable uncertainties, and more work, involving national experts, will be necessary to obtain a verified and generally accepted European data base to estimate the potential for further reductions of fine particles in Europe.

The present implementation (version 2.00) of the RAINS PM module on the Internet (http://www.iiasa.ac.at/rains/Rains-online.html) provides free access to the input data and results to facilitate interaction with national experts.

(11)

Modelling Particulate Emissions in Europe

A Framework to Estimate Reduction Potential and Control Costs

1 Introduction

There is growing concern related to the health effects of fine particles. Recent studies have demonstrated a consistent association between the concentrations of fine particulate matter (PM) in the air and adverse effects on human health (respiratory symptoms, morbidity and mortality) for concentrations commonly encountered in Europe and North America.

Airborne suspended particulate matter can be either primary or secondary in nature. Primary particles (PM) are emitted directly into the atmosphere by natural and/or anthropogenic processes whereas secondary particles are predominantly man-made in origin and are formed in the atmosphere from the oxidation and subsequent reactions of sulfur dioxide, nitrogen oxides, ammonia and volatile organic compounds.

Strategies for controlling particle concentrations in ambient air have to take into account their different origins and address the control potentials for the various sources in a targeted way.

However, to strike a balance among control measures for various pollutants in different economic sectors in several countries is a demanding task, and a large body of information needs to be considered.

Integrated assessment models have been used in the past to identify least-cost strategies that can control multiple precursor emissions leading to acidification, eutrophication and ground-level ozone (Amann and Lutz, 2000). Johansson et al. (2000) have presented an initial attempt to extend the existing framework of the RAINS [Regional Air Pollution Information and Simulation, developed at the International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria] model to address control strategies for fine particulate matter.

The objective of this paper is to present a methodology for estimating primary PM emissions in Europe and the costs involved in reducing primary PM emissions from the various sources in European countries. The remainder of this introductory section reviews the context in which the emission and cost estimates should serve. Section 2 introduces the methodology for estimating emissions and explores the appropriate level of aggregation for a Europe-wide analysis.

Section 3 reviews the available literature sources for the individual source categories and outlines how emission factors were derived for the RAINS model. Cost calculations are the subject of Section 4. Provisional results from the analysis are presented in Section 5, and conclusions are drawn in Section 6. Annex I provides a glossary of frequently used terms.

(12)

1.1 An Integrated Assessment Model for Fine Particulate Matter

Over the last few years, the RAINS model has been used to address cost-effective emission control strategies in a multi-pollutant/multi-effect framework. For this purpose, the RAINS model now includes the control of SO2, NOx, VOC and NH3 emissions as precursors for acidification, eutrophication and ground-level ozone.

For fine particulate matter (PM) there is evidence that several emission sources contribute via various pathways to the concentrations in ambient air. While a certain fraction of fine particles found in the ambient air originates directly from the emissions of those substances (the “primary particles”), a second fraction is formed through secondary processes in the atmosphere from precursor emissions, involving SO2, NOx, VOC and NH3.

Consequently, the search for cost-effective solutions to control the ambient levels of fine particles should balance emission controls over the sources of primary emissions as well as over the precursors of secondary aerosols. Thus, the control problem can be seen as an extension of the “multi-pollutant/multi-effect” concept applied for acidification, eutrophication and ground- level ozone (Table 1.1).

Table 1.1: Air quality management as a multi-pollutant, multi-effect problem.

SO2 NOx NH3 VOC Primary PM

emissions

Acidification √ √ √

Eutrophication √ √

Ground-level ozone √ √

√ √ √ √

Health damage due

to fine particles via secondary aerosols √

Further, a more sophisticated assessment framework could be used for more than just balancing measures for the five pollutants to control fine particles. Such a framework could consider the possible policy objectives for fine particles together with targets for acidification, eutrophication and ground-level ozone, and thereby search for least-cost solutions to address all four environmental problems simultaneously.

The present implementation of the RAINS model contains modules to describe emissions and emission control costs for the first four pollutants. The atmospheric dispersion models employed by RAINS also include the processes leading to the formation of secondary aerosols. Additional modules are necessary to capture primary emissions, control potential and control costs for fine particles, the dispersion of fine particles in the atmosphere and the formation of secondary aerosols from the “conventional" precursor emissions. A module has been developed to assess the health impacts resulting from a certain emission control strategy.

The conceptual extension of the present structure of the RAINS model is illustrated in Figure 1.1, where the additional elements required for the analysis of fine particulate matter are highlighted (Johansson et al., 2000).

(13)

Environmental impacts Economic

activities

Emission control policies

Agriculture

NO_x emissions SO₂ emissions

Solvents, fuels, industry Energy use

NH₃ dispersion

S dispersion

VOC emissions NH₃ emissions

Transport

Critical loads f. acidification

Critical loads f.

eutrophication NO_x dispersion

O₃ formation NH₃ control

& costs

NO_x/VOC control&costs

VOC control

& costs

Emission control costs

Critical levels for ozone

Environmental targets

Primary PM dispersion Other activities PM control

& costs

Primary PM emissions

Secondary aerosols

PM Population exposure SO₂ control

& costs

NO_x control

& costs

O₃ Population exposure

Figure 1.1: Flowchart of the extended RAINS model to address particulate matter.

1.2 The Objectives of an Emission Control Cost Module within the Framework of an Integrated Assessment Model

A central objective of integrated assessment models is to assist in the cost-effective allocation of emission reduction measures across various pollutants, several countries and different economic sectors. Obviously, this task requires consistent information about the costs of emission control at the individual sources, and it is the central objective of this cost module to provide such information.

(14)

control costs among countries, economic sectors and pollutants. Such differences are usually caused, inter alia, by variations in the composition of the various emission sources, the state of technological development and the extent to which emission control measures are already applied.

In order to capture these differences across Europe in a systematic way, a methodology has been developed to estimate the emissions and emission control costs of standard technologies under the specific conditions characteristic for the various European countries. Given the basic assumption of the general availability of control technologies with equal technical properties and costs, a number of country-specific circumstances (level of technological advancement, installation size distribution, labor costs, etc.) are used to estimate the costs for the actual operation of pollution control equipment.

1.3 Summary of Changes Introduced since the Last Release of the RAINS PM Module

This report documents changes that have been introduced in the RAINS PM module since summer 2001 and, consequently, it is an update and extension of the previous report by Lükewille et al., 2001. This section provides a brief summary of the changes.

New sectors

The RAINS model structure has been modified and a number of new emission categories have been introduced, including several industrial processes, mining, storage and handling of bulk materials, open burning of agricultural and residential waste, construction, and other miscellaneous sources (cigarette smoking, barbeques, etc.). A full list of sectors distinguished in RAINS can be found in Table 2.2, Table 2.4, Table 2.5, and Table 2.6.

Revisions

Several emission categories have been revised, i.e., updates of emission factors, activity data and removal efficiencies, and structure modifications within relevant sectors were carried out.

For stationary combustion sources, significant changes in the assumptions about the size fraction distribution of particulate emissions were introduced, as well as an update of size- fraction specific removal efficiencies. Additionally, emission factors for biomass combustion are no longer estimated on the basis of ash content but are derived instead from the literature. A major structural change was introduced for the residential sector for solid fuel combustion, where, instead of one category, RAINS now distinguishes between fireplaces, stoves, single- family house boilers, and medium size boilers. For the latter two, a distinction between manual and automatic fuel loading installations is made.

Within the industrial processes category, the iron and steel sector has been extended to distinguish between sinter plants, pig iron, open hearth, basic oxygen, and electric arc furnace

(15)

iron and steel foundries. Additionally, fugitive emissions from the iron and steel industry are modeled separately.

The transport sector structure has been extended: motorbikes are treated separately and there are now separate off-road categories for construction and industry, agriculture, rail, inland navigation, shipping, and other. In addition, emission factors for vehicles with spark ignition engines were updated.

Based on new information available for the agricultural sector, the structure of the sector was modified to include “arable farming” in the list of sub-sectors. New emission factors for livestock housing were introduced and the set of control techniques was updated.

New fuels

Recognizing the fact that alternative fuels might play an important role in the near future, a number of new fuel categories were distinguished including methanol, ethanol, and hydrogen. A full list of fuels can be found in Table 2.3.

New control options

Modifications and extensions of the model sectoral structure required the introduction of new control options. In some cases these are, technically speaking, the same options as in the previous model version, e.g., electrostatic precipitator, fabric filters, etc., but applicable specifically to industrial processes and, therefore, their removal efficiency and cost characteristics might be different. For several sectors where fugitive emissions play an important role, options to control these losses were added. A few new abatement options were added for the transport sector, e.g., PSA particulate filter. The list of options was also extended for agriculture. A complete list of abatement techniques, together with assumed reduction efficiencies, is provided in Table 2.7, Table 2.8, Table 2.9, Table 2.13, Table 2.14, Table 2.15.

Cost data

The cost data were revised and further developed. To facilitate transparency of the method applied, some examples of how costs were calculated are provided in Chapter 4.

New model features

The model provides several new features that allow for easier viewing of input data, the assumptions made for several parameters, and output. Specifically, the user can display emission factors in either standard RAINS units, e.g., g/MJ for energy use sectors, or as flue gas concentrations for stationary combustion sources, i.e., mg/m³, and g/km or g/kWh for transport categories. This makes it easier to compare the model emission factors (controlled and uncontrolled) with measurement data and legislation.

The Internet version of the RAINS PM module has been updated and is available from the

(16)

(17)

2 A Module to Estimate Emissions of Fine Particulate Matter

2.1 Methodology

The emissions of particulate matter (PM) in the RAINS model are calculated for three different size classes:

• fine fraction (PM2.5),

• coarse fraction (PM10 - PM2.5) and

• large particles (PM_>10 µm).

Thereby, PM10 is calculated as the sum of fine and coarse fractions and total suspended particles (TSP) as the sum of fine, coarse and PM_>10 fractions.

The methodology includes the following three steps:

• In a first step, country-, sector- and fuel-specific “raw gas” emission factors for total suspended particles (TSP) are derived:

 For solid fuels (excluding biomass and use of solid fuels in small residential installations) the mass balance approach is used where ash content (ac) and heat value (hv) of fuels and ash retention in boilers (ar) are considered:

efTSP = ac/hv * (1 – ar)

 For liquid fuels, biomass, solid fuels used in small residential installations, industrial processes, mining, storage and handling of bulk materials, waste incineration, agriculture¹, and transport, TSP emission factors are taken from the literature.

• In a second step, “raw gas” emission factors for each of the size fractions are estimated.

This is done based on size fraction profiles reported in the literature for a variety of installations. They are typically given for PM10 and PM2.5 and are fuel- and installation (sector)-specific. The typical profiles are applied to the country-, fuel- and sector-specific

“raw gas” TSP emission rates (see first step) to derive the size-specific emission factors used in RAINS.

• In a third step, actual PM emissions are calculated for the three size fractions. For a given country (i), PM emissions of size fraction (y) are calculated by applying a general formula across every fuel (activity) and sector, taking into account the application rates of control technologies and size fraction specific emission removal efficiencies,

(18)

∑

∑ ⁼ ⁻

=

m k j

m k j i y m y

k j i k j i m

k j

y m k j i y

i

E A ef eff X

E

, ,

, , , , ,

,

, , , ,

,

( 1 )

(1)

where:

i,j,k,m Country, sector, fuel, abatement technology;

y Size fraction, i.e. fine, coarse, PM_>10;

Ei,y Emissions of PM in country i for size fraction y;

A Activity in a given sector, e.g. coal consumption in power plants;

ef “Raw gas” emission factor;

effm,y Reduction efficiency of the abatement option m for size class y, and;

X Actual implementation rate of the considered abatement, e.g., percent of total coal used in power plants that are equipped with electrostatic precipitators.

If no emission controls are applied, the abatement efficiency equals zero (effm,y = 0) and the application rate is one (X = 1). In that case, the emission calculation is reduced to simple multiplication of activity rate by the “raw gas” emission factor.

2.2 Aggregation of Emission Sources

Emissions of PM are released from a large variety of sources with significant technical and economic differences. Conventional emission inventory systems, such as the CORINAIR inventory of the European Environmental Agency, distinguish more than 300 different processes causing various types of emissions.

In the ideal case, the assessment of the potential and costs for reducing emissions should be carried out at the very detailed process level. In reality, however, the necessity to assess abatement costs for all countries in Europe, as well as focus on emission levels in 10 to 20 years from now, restricts the level of detail which can be maintained. While technical details can be best reflected for individual (reference) processes, the accuracy of estimates on an aggregated national level for future years will be seriously hampered by a general lack of reliable projections of many of these process-related parameters (such as future activity rates, autonomous technological progress, etc.). For an integrated assessment model focusing on the pan-European scale it is therefore imperative to aim at a reasonable balance between the level of technical detail and the availability of meaningful data describing future development, and to restrict the system to a manageable number of source categories and abatement options.

2.2.1 Criteria for Aggregations

For the RAINS PM module, an attempt was made to aggregate the emission producing processes into a reasonable number of groups with similar technical and economic properties.

Considering the intended purposes of integrated assessment, the major criteria for aggregation were:

(19)

The importance of the emission source. It was decided to target source categories with a contribution of at least 0.5 percent to the total anthropogenic emissions in a particular country.

The possibility of defining uniform activity rates and emission factors.

The possibility of constructing plausible forecasts of future activity levels. Since the emphasis of the cost estimates in the RAINS model is on future years, it is crucial that reasonable projections of the activity rates can be constructed or derived.

The availability and applicability of “similar” control technologies.

The availability of relevant data. Successful implementation of the module will only be possible if the required data are available.

It is important to define carefully the appropriate activity units. They must be detailed enough to provide meaningful surrogate indicators for the actual operation of a variety of different technical processes, and aggregated enough to allow a meaningful projection of their future development with a reasonable set of general assumptions. As explained later in the text, some of the RAINS sectors contain a number of PM emitting processes. It is often the case that for such aggregated sectors some emission control options are not necessarily applicable to all processes (emission sources) that are represented by the activity.

Table 2.1 presents major sectors included in the RAINS PM module and their contribution to total European PM emissions that are estimated in this study for 1995. The RAINS source structure shown distinguishes ten emission categories for mobile sources and three for stationary combustion sources that are split by relevant fuels (see Table 2.2), and 17 other sectors. Some categories are further disaggregated to distinguish, for example, between existing and new installations in power plants, or between tire and brake wear for non-exhaust emissions from transport (for a full list of RAINS sectors see Table 2.3, Table 2.4, Table 2.5).

The sectoral structure of the RAINS model is not directly compatible with that of CORINAIR or the UNECE reporting standard (NFR – Nomenclature For Reporting) (UNECE, 2002).

Tables presented in this section provide a broad reference to the CORINAIR SNAP’94 and UNECE-NFR categories. In several cases the relation can be established only for a primary sector, i.e., the sum of all RAINS categories for power and district heating plants can only be compared with the sum of several SNAP entries. RAINS contains a feature to aggregate/display emissions into the CORINAIR SNAP level 1 as well as NFR level 1 and 2.

The following sections define the source categories distinguished in the RAINS model in more detail and provide the corresponding SNAP source sectors of the CORINAIR inventory as well as the UNECE-NFR categories.

2.2.2 Stationary Combustion Sources

(20)

similar share of emissions and together they represent nearly 65 percent of the total (Table 2.1).

An attempt has been made to design an emission source structure that represents the most important sources and factors influencing emissions of PM. The following tables present the RAINS model sectors used in the PM calculation; for the most part they are compatible with the structure of the other RAINS modules although new elements are introduced. More details are given in Section 3.

Table 2.1: Major sectors included in the RAINS PM module and their contribution to total European PM emissions in 1995 as estimated in this study.

RAINS sector Emissions [kt]

Share of total European emissions in

1995 [%]

Primary Secondary TSP PM10 PM2.5 TSP PM10 PM2.5

Stationary Power plants 1410 785 378 13.4 15.5 11.9 combustion Industrial combustion 419 182 87 4.0 3.6 2.8

Domestic combustion 3057 993 544 29.1 19.6 17.2

Process Pig iron 287 42 28 2.7 0.8 0.9

emissions Sinter and pellets 277 63 34 2.6 1.2 1.1 Basic oxygen furnaces 325 291 244 3.1 5.7 7.7 Electric arc furnaces 103 86 73 1.0 1.7 2.3 Other Iron and Steel 430 368 279 4.1 7.3 8.8

Non-ferrous metals 66 57 48 0.6 1.1 1.5 Cement and lime 283 200 144 2.7 3.9 4.5 Other processes 510 261 154 4.9 5.1 4.9

Mining 113 57 6 1.1 1.1 0.2

Storage and Industrial products 399 181 18 3.8 3.6 0.6 handling Agricultural products 65 18 3 0.6 0.3 0.1 Road transport Heavy duty vehicles 185 182 179 1.8 3.6 5.6

Light duty vehicles 234 231 220 2.2 4.5 6.9 Motorcycles, mopeds 13 12 11 0.1 0.2 0.4

Non-exhaust 462 93 30 4.4 1.8 1.0

Off-road Construction and Industry 32 31 29 0.3 0.6 0.9

transport Agriculture 135 128 121 1.3 2.5 3.8

Rail 34 32 30 0.3 0.6 1.0

Inland waterways 29 27 26 0.3 0.5 0.8

Other land-based 23 20 18 0.2 0.4 0.6

Maritime activities 141 134 127 1.3 2.6 4.0 Open burning of waste 265 265 200 181 2.5 3.9

Agriculture Livestock 492 221 45 4.7 4.4 1.4

Other 511 28 0 4.9 0.6 0.0

Other sources Construction dust 83 41 4 0.8 0.8 0.1

Residential ⁽¹⁾ 87 87 87 0.8 1.7 2.8

Other 26 21 17 0.2 0.4 0.5

TOTAL 10498 5072 3167 100.0 100.0 100.0

(1) Food preparation, barbeques, cigarette smoking, and fireworks

(21)

Table 2.2: RAINS sectors related to stationary sources with energy combustion.

RAINS sector RAINS code NFR

category

SNAP sector Road transport, Tire wear

Heavy duty vehicles (trucks, buses and others) TRT_RD_HD Motorcycles, four-stroke TRT_RD_M4 Motorcycles and mopeds (also cars), two-stroke TRT_RD_LD2 Light duty cars and vans, four-stroke TRT_RD_LD4 Light duty cars, four-stroke, gasoline direct injection TRT_RDXLD4

1A3b 0707

Road transport, brake wear

Heavy duty vehicles (trucks, buses and others) TRB_RD_HD Motorcycles, four-stroke TRB_RD_M4 Motorcycles and mopeds (also cars), two-stroke TRB_RD_LD2 Light duty cars and vans, four-stroke TRB_RD_LD4 Light duty cars, four-stroke, gasoline direct injection TRB_RDXLD4

1A3b 0707

Road transport, abrasion of paved roads

Heavy duty vehicles (trucks, buses and others) TRD_RD_HD

Motorcycles, four-stroke TRD_RD_M4

Motorcycles and mopeds (also cars), two-stroke TRD_RD_LD2

Light duty cars and vans, four-stroke TRD_RD_LD4

Light duty cars, four-stroke, gasoline direct injection TRD_RDXLD4

1A3b

(25)

2.3 Emission Factors

Emission factors are the key to assess PM emissions accurately. For the present study it has been decided to identify, as far as possible, the main factors that could lead, for a given source category, to justified differences in emission factors across countries. The aim has been to collect country-specific information to quantify such justifiable deviations from values reported in the general literature. When this was not possible or when a source category makes only a minor contribution to total emissions, emission factors from the literature were used.

Within the PM module, unabated emission factors of total suspended matter (TSP) are the basis for deriving emission factors for fractions of the total range of PM mass concentrations.

Emission factors of fine PM for two size classes, PM10 (ø < 10 µm) and PM2.5 (ø < 2.5 µm), are calculated from the TSP estimates by using typical (source-specific) size profiles available in the literature.

2.3.1 Emission Factors for Stationary Combustion Sources

Due to the large overall contribution of the stationary combustion of solid fuels to total PM emissions (varying between 50 and 65 percent for PM2.5 and TSP), an attempt has been made to derive country-specific emission factors for power plants, industrial boilers, waste processing plants and domestic ovens. Emission factors have been computed by applying a mass balance approach. Country-specific information on the ash contents of different fuels (IEA, 1998), heat values (RAINS database), and the fraction of ash retained in the respective boiler type was used (e.g., Kakareka et al., 1999; EPA, 1998a) (compare Equation 2). Emission factors for total suspended particulate matter (TSP) are estimated in a first step:

efTSP = ac/hv * (1 - ar)*10 (2)

where:

ef unabated emission factor [g/MJ], ac ash content [%],

hv lower heat value [GJ/t], ar fraction of ash retained in boiler .

In a second step, the emissions of fine particulate matter (for two size fractions: PM10 and PM2.5) are calculated from the TSP estimates by using typical size profiles available in the literature (e.g., Ahuja et al., 1989; Houck et al., 1989; EPA, 1998a; AWMA, 2000; Kakareka et al., 1999). The order of magnitude of the emission factors obtained with this method was checked against values reported in the literature, e.g., TA Luft, 1986; Soud, 1995, and summarized by Dreiseidler et al. (1999).

For PM emissions from the combustion of liquid fuels (gasoline, diesel, heavy fuel oil), natural gas, biomass, and solid fuels burned in small residential installations emission factors from the

(26)

2.3.2 Emission Factors for Mobile Sources

For on-road mobile sources, RAINS derives emission factors from the studies carried out in connection with the Auto Oil 1 and 2 Programmes (EC, 1999). For gasoline vehicles, additionally the following studies were used: Hildemann et al., 1991; Norbeck et al., 1998a;

Durbin et al., 1999; Kwon et al, 1999; CONCAWE, 1998 (see Section 3.8.1.3). Thus, the emission factors used in RAINS for the various vehicle categories are based on the full range of country-specific factors such as driving pattern, fleet composition, climatic conditions, etc. that was considered in the Auto Oil analyses. For the RAINS assessment, fuel-related emission factors for diesel vehicles were obtained by dividing the volume of PM emissions calculated in the Auto Oil project for the RAINS vehicle categories by the respective fuel consumption.

For off-road sources, a range of American and European studies were used, e.g., EPA, 1991;

BUWAL, 2000a; Breadsley et al., 1998; Norbeck et al., 1998ab; Kean et al., 2000; and specifically for shipping: Lloyd’s Register, 1997 and Wright, 1997, 2000 (for details see Section 3.8.1.4).

Non-exhaust emission factors for road transport were extracted from various literature sources (see Section 3.8.2). Since such emission factors are usually reported in grams per kilometer (g/km), the fuel-efficiencies of the various vehicle categories have been used to convert them into the fuel-related emission factors. Time-dependent and country-specific fuel efficiencies are taken from the studies conducted for the Auto/Oil 2 Programme (EC, 1999). Although highly uncertain, the RAINS model treats emissions from tire lining wear, brake wear and abrasion of paved roads as separate sources (see Sections 3.8.2.1, 3.8.2.2, 3.8.2.3).

2.3.3 Emission Factors for Other Sources

The RAINS model includes a long list of non-combustion emission sources (Table 2.4). Here, only major categories and primary sources of emission factor data will be addressed. More detailed information can be found in respective sections of this document and listed literature.

Emission rates for the iron and steel industry and non-ferrous metal industry are based primarily on EPA, 1998a; Rentz et al., 1996; TA Luft, 1986; AWMA, 2000; UBA, 1998a; and a review by Passant et al. (2000). For agriculture, two major studies are used, i.e., Takai et al., 1998 and ICC & SRI, 2000. Information on particulate emissions and emission rates for the remaining sectors, i.e., mining, storage and handling of bulk materials in industry and agriculture, open burning of waste, construction activities, and other miscellaneous sources, is scarce. The recently completed project CEPMEIP (Co-ordinated European Programme on Particulate Matter Emission Inventories, Projections and Guidance) (CEPMEIP, 2002) proved very helpful in compiling this information. Additionally, reports from EPA (1995, 1998a), Dreiseidler et al.(1999), Ecker and Winter (2000), Schindler and Ronner (2000), Staubenvoll and Schindler (1998) and Berdowski et al. (1997) were used.

(27)

2.4 Emission Control Options

2.4.1 Stationary Sources

In addition to the obvious “structural changes” that lead to a lower consumption of emission generating fuels, there are several end-of-pipe options for reducing particulate matter emissions from stationary sources, e.g., Darcovich et al., 1997; Soud, 1995; TA Luft, 1986; Rentz et al., 1996). The following paragraphs briefly review the main options and their technical characteristics.

2.4.1.1 A Review of Available Control Options

Inertial Settlers and Cyclones

The general principle of cyclones is the inertial separation of particles from the gas stream.

Particulate-laden gas is forced to change direction, and the inertia of the particles causes them to continue in the original direction. In Western Europe multi-cyclones are usually only used as pre-dedusters (pre-cleaners) for the collection of medium-sized and coarse particles. The net downward motion of particles will arise at sizes larger than 5 µm. Thus gravity settling will be efficient only on large particles (40 to 50 µm). The removal efficiency drops if the fines content of the particulate matter is significant and generally does not lead to a substantial reduction of PM0.1 emissions.

Wet Scrubbers

In the most widely used Venturi scrubber, water is injected into the flue gas stream at the Venturi throat to form droplets. Fly ash particles impact with the droplets forming a wet by- product, which then generally requires disposal. The process can also have a high energy consumption due to the use of sorbent slurry pumps and fans.

The efficiency of wet scrubbing for particulate removal depends on the particle size distribution.

The system efficiency is reduced as the particle size decreases.

Fabric Filters

Dust particles moving through fabric filters often form a porous cake on the surface of the fabric. This cake normally does the bulk of the filtration. Conventional reverse-gas-cleaned fabric filters (baghouses, RGB) are being quickly replaced by pulse-jet fabric filters (PJFF).

Periodic short, powerful bursts of air are used to clean the fabric mounted in cylindrical bags.

Interception (fibrous or granular filter media) is effective on particles down to 2-3 µm. Effective processes to remove particles smaller than 0.2 µm are thermal precipitation (cold collection system) and diffusional deposition (fibrous or granular filter media and small liquid droplets).

(28)

Electrostatic Precipitators (ESP)

In electrostatic precipitators (ESP), particles are given an electric charge by forcing them through a region in which gaseous ions flow. Electrodes in the center of the flow channel maintain a high voltage, forcing particles to move out of the flowing gas stream onto collector plates. The particles are removed from the plates by knocking them loose or by washing with water. Developments in ESP technology aim especially at improving the collection of ultra-fine particles. ESP can tolerate temperatures as high as 400 ^oC.

The performance of fabric filters and some scrubbers can also be enhanced with electrostatic charging. Electrostatic force is the strongest process commonly used as PM removal technology that can act on fine particles smaller than 2-3 µm.

High Temperature, High Pressure (HTHP) Particulate Control

During the last decade there have been significant advances towards the commercialization of combined cycle systems, such as the integrated gasification combined cycle (IGCC) and pressurized fluidized bed combined cycle (PFBCC). Commercial- and demonstration-scale designs are currently used for power generation in the United States, Europe and Japan. An important component in combined cycle power systems is a high temperature, high pressure (HTHP) particulate control device.

Efficient hot gas particulate filtration is necessary to protect the downstream heat exchanger and gas turbine components from fouling and erosion to meet emission requirements. A range of technologies has been proposed for hot gas particulate filtration but few have been developed sufficiently to enable commercial exploitation in combined cycle power systems.

2.4.1.2 Control Options Implemented in the RAINS Model

In the interest of keeping a European-scale analysis manageable, the RAINS model considers a limited number of emission control options reflecting groups of technological solutions with similar emission control efficiencies and costs. For large boilers in industry and power stations, and industrial processes the following options are available:

Cyclones;

Wet scrubbers;

Electrostatic precipitators (three stages, i.e., one field, two fields, and more than two fields);

Wet electrostatic precipitators;

Fabric filters;

Regular maintenance of oil fired industrial boilers;

Two stages (low and high efficiency) of fugitive emissions control measures.

These options are divided into three categories, i.e. power plants, industrial combustion, and industrial processes that can have different emission reduction and cost characteristics. The actual choice of options for a given sector is made on the basis of reviews of real-life applications (e.g., TA Luft, 1986; AWMA, 2000), information from industrial sources and

(29)

environmental agencies, e.g., Umwelbundesamt (UBA, 1998a). The RAINS model considers size-fraction specific removal efficiencies for these control options (Table 2.7).

Table 2.7: Size-fraction specific removal efficiencies for abatement options used in RAINS for power plants and industry.

Removal efficiency Control technology RAINS code

> PM10 Coarse Fine

Cyclone CYC, _CYC 90 % 70 % 30 %

Wet scrubber WSCRB, _WSCRB 99.9 % 99 % 96 %

Electrostatic precipitator, 1 field ESP1, _ESP1 97 % 95 % 93 % Electrostatic precipitator, 2 fields ESP2, _ESP2 99.9 % 99 % 96 % Electrostatic precipitator, 3 fields and more ESP3P, _ESP3P 99.95 % 99.9 % 99 % Wet electrostatic precipitator PR_WESP 99.95 % 99.9 % 99 %

Fabric filters FF, _FF 99.98 % 99.9 % 99 %

Regular maintenance, oil fired boilers GHIND 30 % 30 % 30 % Good practice (industrial processes –

fugitive), stage 1 PRF_GP1 20 % 15 % 10 %

Good practice (industrial processes –

fugitive), stage 2 PRF_GP2 75 % 50 % 30 %

For small and medium size boilers in the residential/commercial sector, a number of measures, depending on the size, fuel, and operation mode (manual or automatic loading), are available:

Cyclones;

Fabric filters;

Regular maintenance of oil fired boilers;

New type of boiler, e.g., pellets or wood chips.

For domestic sources, i.e., fireplaces, single-family boilers, the principal option is a switch to a newer type of installation. Additionally for fireplaces, an option of installing a catalyst or non- catalyst insert is included. Modernization options (two stages potentially including catalytic and non-catalytic and/or primary and secondary air deflectors) are included for coal and wood stoves. The data on efficiencies (Table 2.8) and costs of these options for wood burning originates from Houck and Tiegs (1998). This study refers to the American situation and the data need to be reviewed taking into account European conditions. At this stage, however, no similar data for Europe could be found. Techniques to control emissions from coal burning installations are primarily “placeholders” that can be used when more information about possibilities to control these sources is available. As with other categories, regular maintenance of oil-fired boilers is also included. Size-fraction specific removal efficiencies for these control options are given in Table 2.8.

(30)

Table 2.8: Size-fraction specific removal efficiencies for abatement options used in RAINS for residential combustion sources.

Removal efficiency Control technology RAINS code

> PM10 Coarse Fine Fireplaces, non-catalytic insert FP_ENC 44 % 44 % 44 % Fireplaces, catalytic insert FP_CAT 47 % 47 % 47 % New domestic stoves (coal), stage 1 COAL1 30 % 30 % 30 % New domestic stoves (coal), stage 2 COAL2 50 % 50 % 50 % New domestic boilers (coal) NB_COAL 40 % 40 % 40 % New domestic stoves (wood), non-

catalytic WOOD1 63 % 63 % 63 %

New domestic stoves (wood), catalytic WOOD2 65 % 65 % 65 % New wood boilers (wood chips, pellets) MB_PELL 89 % 89 % 89 % Regular maintenance, oil fired boilers GHDOM 30 % 30 % 30 %

Cyclone MB_CYC 90 % 70 % 30 %

Fabric filters MB_BAG, _PLBAG 99.98 % 99.9 % 99 %

For several non-combustion PM sources included in the model, a range of control options is included. It has to be noted, however, that information on their removal efficiencies as well as costs is very scarce or not available at all. The only sector for which more extensive discussion of control options is available is agriculture (Takai et al., 1998; ICC &SRI, 2000). Assumptions made in RAINS on removal efficiency for the included options are summarized in Table 2.9.

Table 2.9: Size-fraction specific removal efficiencies for abatement options used in RAINS for non-combustion sources.

Removal efficiency

Control technology RAINS code

> PM10 Coarse Fine Agriculture

Feed modification (all livestock) FEED_MOD 45 % 35 % 10 % Hay-silage for cattle HAY_SIL 70 % 40 % 10 %

Free range poultry FREE 40 % 15 % 5 %

Low-till farming, alternative cereal harvesting ALTER 40 % 15 % 5 % Good practice (other animals) [generic option] AGR1 40 % 15 % 5 % Other sources

Good practice, storage and handling STH_GP 50 % 20 % 10 % Good practice in oil and gas industry, flaring FLR_GP 40 % 15 % 5 % Ban on open burning of waste BAN 100 % 100 % 100 % Good practice in mining industry MINE_GP 55 % 47 % 25 % Spraying water at construction sites SPRAY 50 % 20 % 10 % Filters in households (kitchen) FILTER 50 % 20 % 10 % Generic, e.g., street washing RESP1 n.d.⁽¹⁾ n.d. n.d.

(1) not defined yet

(31)

2.4.2 Mobile Sources

Primary particle emissions from mobile sources have two entirely different origins: exhaust due to fuel combustion; and non-exhaust emissions, i.e., tire and brake wear and road abrasion or re- suspension (dust swept up or entrained into the air by passing traffic). In this section options to control exhaust emissions of PM, as well as their implementation in RAINS, are discussed.

2.4.2.1 A Review of Available Control Options

Emission control options for mobile sources can be divided into the following categories:

Changes in fuel quality, e.g., decreases in sulfur content. Changes in fuel specifications may provide engine manufactures with greater flexibility to use new emission reduction technologies.

Changes in engine design, which result in better control of the combustion processes in the engine.

Flue gas post-combustion treatment, using various types of trap concepts and catalysts to convert or capture emissions before they leave the exhaust pipe.

Better inspection and maintenance. Examples are: in-use compliance testing, in-service inspection and maintenance, on-board diagnostic systems.

Diesel Fuels and Clean Diesel Engines

High sulfur or aromatics contents have an impact on the quantity and quality of particulate matter emissions. They also interfere with several technologies controlling diesel exhaust. A reduction of fuel density lowers NO_x and PM emissions, but on the other hand it increases hydrocarbon (HC) and carbon monoxide (CO) exhaust. The use of synthetic diesel fuel, gained from feedstock such as gas or coal, significantly reduces all pollutant emissions, including PM.

Other measures, which may result in lower PM emissions, are the use of bio-diesel, derived from various vegetable oils, and of dimethyl ether (DME), made, for example, from natural gas and coal (http://www.dieselnet.com).

Changes in diesel engine design have reduced emissions from diesel vehicles by more than 90 percent. Important improvements are electronic controls and fuel injectors to deliver fuel at the best combination of injection pressure, injection timing and spray location, air-intake improvements, combustion chamber modifications, exhaust gas re-circulation and ceramic in- cylinder coatings (see also Cofala and Syri, 1998b).

Diesel Catalyst Technology

Catalysts increase the rate of chemical reaction. In emission control applications heterogeneous catalysts are used, which are supported on high surface area porous oxides. Two processes may cause malfunction of emission control catalysts: poisoning and thermal deactivation. The

(32)

Diesel oxidation catalysts were first introduced in the 1970s in underground mining as a measure to control CO. Today catalysts are used on many diesel cars in Europe, primarily to control PM and hydrocarbon emissions. Early diesel catalysts utilized active oxidation formulations such as platinum on alumina. They were very effective in oxidizing emissions of CO and HC as well as the organic fraction (SOF) of diesel particles.

However, catalysts also oxidize sulfur dioxide, which is present in diesel exhaust from the combustion of sulfur-containing fuels. The oxidation of sulfur to SO2 leads to the generation of sulfate particulate matter. This may significantly increase total primary particle emissions, although the SOF PM fraction is reduced. Newer diesel oxidation catalysts are designed to be selective, i.e., to obtain a compromise between sufficiently high HC and SOF activity and acceptably low formation of SO2.

Diesel Particulate Traps

Diesel particulate traps physically capture diesel particles preventing their release to the atmosphere. Diesel traps work primarily through a combination of deep-bed filtration mechanisms, such as diffusional and inertial particle deposition. The most common filter materials are ceramic wall-flow monoliths and filters made of continuous ceramic fibers. A number of methods have been proposed to regenerate diesel filters.

Passive filter systems utilize a catalyst to lower the soot combustion temperature. Active filter systems incorporate electric heaters or fuel burners to burn the collected particles.

The regeneration of a diesel filter is characterized by a dynamic equilibrium between the soot being captured in the filter and the soot being oxidized. The rate of soot oxidation depends on the filter temperature. At temperatures that are typically found in diesel exhaust gases, the rate of soot oxidation is small. Therefore, to facilitate filter regeneration, either the exhaust gas temperature has to be increased or a catalyst has to be applied. The catalyst can be applied directly onto the filter media or dissolved in the fuel as a fuel additive.

Wall-flow monoliths became the most popular diesel filter design. They are derived from flow- through catalyst supports where channel ends are alternately plugged to force the gas flow through porous walls acting as filters. The monoliths are made of specialized ceramic materials.

Most catalyzed diesel traps utilize monolithic wall-flow substrates coated with a catalyst. The catalyst lowers the soot combustion temperature, allowing the filter to self-regenerate during periods of high exhaust gas temperature. Filters of different sizes, with and without catalysts, have been developed and are available as standard products.

The CRT (Continuously Regenerating Trap) system for diesel particulate utilizes a ceramic wall-flow filter to trap particles. The trapped PM is continuously oxidized by nitrogen dioxide generated in an oxidation catalyst, which is placed upstream of the filter. The CRT requires practically sulfur-free fuel for proper operation.

Fuel additives (fuel soluble catalysts) can be used in passive diesel trap systems to lower the soot combustion temperature and to facilitate filter regeneration. The most popular additives

(33)

include iron, cerium, copper, and platinum. Many laboratory experiments and field tests have been conducted to evaluate the regeneration of various diesel filter media using additives.

Cerium additive is utilized in a commercial trap system for diesel cars.

Electric regeneration of diesel traps has been attempted in off- and on-board configurations.

On-board regeneration by means of an electric heater puts a significant additional load on the vehicle electrical system. Partial flow layouts or regeneration with hot air are more energy efficient. An on-board, hot air regenerated diesel trap was tested on over 2000 urban buses in the U.S. A system with off-board electric regeneration has also been developed and commercialized.

Diesel fuel burners can be used to increase the exhaust gas temperature upstream of a trap in order to facilitate filter regeneration. Fuel burner filters can be divided into single point systems and full flow systems. The full flow systems can be regenerated during regular vehicle operation but require complex control to ensure a thermally balanced regeneration. An advanced system featuring electronically controlled full flow burner regeneration has been developed.

Diesel soot has microwave absorption properties and there are filter substrate materials that are transparent to microwave irradiation. Microwave heating is another method to regenerate diesel particle filters.

2.4.2.2 Control Options Implemented in the RAINS Model

The options to control vehicle emissions in RAINS simulate the effects of implementation of European legislation on mobile sources. Table 2.10 presents the development of emission standards on diesel light-duty vehicles since 1990. Standards for heavy-duty trucks are presented in Table 2.11. Emission limit values for off-road vehicles are presented in Table 2.12.

The regulations for off-road diesel engines are introduced in two stages: Stage I implemented in 1999 and Stage II implemented from 2001 to 2004, depending on the engine power output.

Emission limit values are similar to EURO I and EURO II standards for heavy-duty vehicles.

The equipment covered by the standard includes industrial drilling rigs, compressors, construction wheel loaders, bulldozers, off-road trucks, highway excavators, forklift trucks, road maintenance equipment, snow plows, ground support equipment in airports, aerial lifts and mobile cranes. Agricultural and forestry tractors have the same emission standards but different implementation dates. Engines used in ships, railway locomotives, aircraft, and generating sets are not covered by the standards.

Modeling Particulate Emissions in Europe. A Framework to Estimate Reduction Potential and Control Costs

Interim Report IR-02-076

Modelling Particulate Emissions in Europe

A Framework to Estimate Reduction Potential and Control Costs

Contents

Acknowledgments

Abstract

Modelling Particulate Emissions in Europe

A Framework to Estimate Reduction Potential and Control Costs

1 Introduction

2 A Module to Estimate Emissions of Fine Particulate Matter

∑

∑ = −

=

E A ef eff X

E

( 1 )

∑ ⁼ ⁻